Self-checkout for supermarket Project and Laboratories on Communication Systems Members: - Asnani Sorath - Carella Giuseppe - Mensio Martino
Objectives Prototype for automatic cash register - identification of user and products - creation and saving of receipts - management of the inventory and shelf-refilling
System overview
Client side
Steps for executing a purchase
User authentication
Identification of products ● The user takes the picture of the barcode. ● The picture is sent to the server. The server sends the information about ● the product.
Product removal features ● Scenario: the user wants to remove “coca cola”, because he has not enough money to buy it.
Proceeding with payment and storing receipt
Error management
User authentication failed ● Bad quality of QR code picture.
Identification of product failed ● Bad quality of bar code picture.
Connection down ● The ethernet cable is disconnected.
Server side
Communication with the board - UDP channel - stateless communication: each request is independent from the others
Code recognition and processing Types of codes: - 1D codes: for the products - 2D codes: for customers Steps: - receive image - Zebra-crossing library to detect code - lookup in the database - provide result back to the board
GUI basic logging features
Interaction with third-party Missing products in the local service database - only for operators - fast insertion https://upcdatabase.org/
Website and database
Database Design
System analysis
Ease-of-use, Applicability and Marketing perspective User-friendly both on customer and administrator side. ● ● A lot of functionalities to execute the purchase. ● Each single step of the purchase is well explained. Automatic shopping in supermarkets with high attendance. ● Speed up the buying phase of customers. ●
Quick estimation of costs
Estimation ● LCD: 80$ ● Tunes module: 10$ ● FEZ Spider II: 60$ ● Button: 5$ ● Camera: 30$ ● Ethernet module: 20$ ● Firmware on the processor: 500$ ● Server Side: 500$ ● WebSite and database: 1000$ ● Assistance: 300$ per year and it also covers damages for two years. ● Total cost: 2505$
Dedicated hardware for code recognition would be much better Conclusions (barcode reader) - fast image → code - fast interaction with server - accuracy
Recommend
More recommend